Provided by

arXiv

arxiv.org
𝕏gh
0.0/5
3 views
0 saved
Freeadvanced
ResearchersAI EngineersData ScientistsStudents
Introduces Generative Adversarial Networks (GANs), a novel framework for estimating generative models via a minimax game.

Overview

This seminal paper introduces Generative Adversarial Networks (GANs), a powerful class of generative models. GANs learn to generate realistic data by pitting two neural networks against each other: a generator that creates fake samples, and a discriminator that tries to distinguish real from fake. This adversarial process drives both networks to improve, resulting in high-quality synthetic data generation across various domains, revolutionizing image synthesis, data augmentation, and more.

Abstract

We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equalling 1/2 everywhere. In the case where G and D are defined by multi-layer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation. Experiments demonstrate the potential of the framework to generate visually plausible samples in several datasets.